Matt Hartman

If you just want to send to trello (not only zo), can send your private context to the trello API

Optional connectivity with zo computer for external skills

We built the app using all Huggingface local models so we could run inference 100% locally and therefore 100% privately on your computer.

Whisper tiny.en is only 75 MB and accurate for transcription itself. We default to small.en and also offer Whisper Small (multilingual) and others so people can pick what's best for their situation.

I tagged Qwen 2.5 here but we actually used Qwen 3.5 family series. Found them to be the best mix of small, fast, and accurate for cleaning up "ums" etc.

I built in Claude Code because I found it to be the easiest fastest way to build the product. Part way through I think Jesse used Codex